# -*- coding: utf-8 -*-
"""
This file contains models of Poissonian fitting routines for qudi based on the lmfit package.
Copyright (c) 2021, the qudi developers. See the AUTHORS.md file at the top-level directory of this
distribution and on <https://github.com/Ulm-IQO/qudi-core/>
This file is part of qudi.
Qudi is free software: you can redistribute it and/or modify it under the terms of
the GNU Lesser General Public License as published by the Free Software Foundation,
either version 3 of the License, or (at your option) any later version.
Qudi is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See the GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License along with qudi.
If not, see <https://www.gnu.org/licenses/>.
"""
__all__ = ('DoublePoissonian', 'Poissonian', 'multiple_poissonian')
import numpy as np
from scipy.special import gammaln, xlogy
from qudi.util.fit_models.model import FitModelBase, estimator
from qudi.util.fit_models.helpers import smooth_data, sort_check_data, estimate_double_peaks
from qudi.util.fit_models.gaussian import multiple_gaussian
[docs]
def multiple_poissonian(x, mus, amplitudes):
""" Mathematical definition of the sum of multiple scaled Poissonian distributions without any
bias.
WARNING: Iterable parameters "mus" and "amplitudes" must have the same length.
Parameters
----------
x : float
The independent variable to calculate the Poissonian.
mus : iterable
Iterable containing center positions (means) for all Poissonians.
amplitudes : iterable
Iterable containing amplitudes for all Poissonians.
Returns
-------
float
The result given x for the sum of scaled Poissonian distributions.
"""
assert len(mus) == len(amplitudes)
# Use the numerically more stable definition of the Poisson probability mass function.
# The standard definition is quickly blowing up for increasing values of x.
#
# Due to the fact that the poissonian distribution is approaching a normal distribution for
# large values of mu, we define a cut-off value of 1e6. If our independent variables x are at
# or above this value, we will switch to calculating a normal distribution. This ensures that
# this function will remain numerically stable for very large values of x and mu.
if min(x) < 1e6:
return sum(np.exp(xlogy(x, mu) - gammaln(x + 1) - mu) for mu, amp in
zip(mus, amplitudes, amplitudes))
else:
return multiple_gaussian(x, mus, np.sqrt(mus), amplitudes)
[docs]
class Poissonian(FitModelBase):
"""
"""
[docs]
def __init__(self, **kwargs):
super().__init__(**kwargs)
self.set_param_hint('offset', value=0, min=-np.inf, max=np.inf)
self.set_param_hint('amplitude', value=1., min=0, max=np.inf)
self.set_param_hint('mu', value=1., min=0, max=np.inf)
@staticmethod
def _model_function(x, offset, mu, amplitude):
return offset + multiple_poissonian(x, (mu,), (amplitude,))
@estimator('default')
def estimate(self, data, x):
data, x = sort_check_data(data, x)
filter_width = max(1, int(round(len(x) / 20)))
data_smoothed, _ = smooth_data(data, filter_width)
# estimate offset and level data
offset = min(data_smoothed)
data_smoothed -= offset
# estimate other parameters
mu = x[np.argmax(data_smoothed)]
amplitude = max(data_smoothed) / self._model_function(mu, 0, mu, 1)
x_spacing = min(abs(np.ediff1d(x)))
x_span = abs(x[-1] - x[0])
data_span = abs(max(data) - min(data))
estimate = self.make_params()
estimate['mu'].set(value=mu,
min=max(x_spacing, min(x) - x_span / 2),
max=min(x_span, max(x) + x_span / 2))
estimate['amplitude'].set(value=amplitude, min=0, max=2 * amplitude)
estimate['offset'].set(value=offset,
min=min(data) - data_span / 2,
max=max(data) + data_span / 2)
return estimate
@estimator('No Offset')
def estimate_no_offset(self, data, x):
estimate = self.estimate(data, x)
estimate['offset'].set(value=0, min=-np.inf, max=np.inf, vary=False)
return estimate
[docs]
class DoublePoissonian(FitModelBase):
"""
"""
[docs]
def __init__(self, **kwargs):
super().__init__(**kwargs)
self.set_param_hint('offset', value=0, min=-np.inf, max=np.inf)
self.set_param_hint('amplitude_1', value=1., min=0, max=np.inf)
self.set_param_hint('amplitude_2', value=1., min=0, max=np.inf)
self.set_param_hint('mu_1', value=1., min=0, max=np.inf)
self.set_param_hint('mu_2', value=2., min=0, max=np.inf)
@staticmethod
def _model_function(x, offset, mu_1, mu_2, amplitude_1, amplitude_2):
return offset + multiple_poissonian(x, (mu_1, mu_2), (amplitude_1, amplitude_2))
@estimator('default')
def estimate(self, data, x):
data, x = sort_check_data(data, x)
data_smoothed, filter_width = smooth_data(data)
# estimate offset and level data
offset = min(data_smoothed)
data_smoothed -= offset
estimate, limits = estimate_double_peaks(data_smoothed, x, filter_width)
params = self.make_params()
params['amplitude_1'].set(value=estimate['height'][0],
min=limits['height'][0][0],
max=limits['height'][0][1])
params['amplitude_2'].set(value=estimate['height'][1],
min=limits['height'][1][0],
max=limits['height'][1][1])
params['center_1'].set(value=estimate['center'][0],
min=limits['center'][0][0],
max=limits['center'][0][1])
params['center_2'].set(value=estimate['center'][1],
min=limits['center'][1][0],
max=limits['center'][1][1])
params['sigma_1'].set(value=estimate['fwhm'][0] / 2.3548,
min=limits['fwhm'][0][0] / 2.3548,
max=limits['fwhm'][0][1] / 2.3548)
params['sigma_2'].set(value=estimate['fwhm'][1] / 2.3548,
min=limits['fwhm'][1][0] / 2.3548,
max=limits['fwhm'][1][1] / 2.3548)
return params
@estimator('No Offset')
def estimate_no_offset(self, data, x):
estimate = self.estimate(data, x)
estimate['offset'].set(value=0, min=-np.inf, max=np.inf, vary=False)
return estimate